Can you login to zookeeper and look under /zkRoot and see the
consumerGroupId node there?  Does it have actual offsets set?

I also assume you don't have any other settings being set on
spoutConfig like :spoutConfig.maxOffsetBehind ?


On Fri, Nov 6, 2015 at 5:29 PM, Birendra Kumar Singh <[email protected]> wrote:
> I am running my topology on a cluster. I have my 2 GCE VMs running. On one
> of the VMs I am running the zookeeper and nimbus. On another I am running
> the supervisor.
> all the values in SpoutConfig contructor are the same in subsequent restarts
>
> spoutConfig = new SpoutConfig(hosts, topic, zkRoot, consumerGroupId); ---
> all values remain the same
>
>
> On Fri, Nov 6, 2015 at 2:50 PM, Stephen Powis <[email protected]> wrote:
>>
>> Are you running your topology in local mode or on a cluster?  What are
>> you using as your consumer id in the spout?
>>
>> IE when you're building your spout config like here:
>>
>> SpoutConfig spoutConfig = new SpoutConfig(zkHosts, topic, zkRoot,
>> zkSpoutId);
>>
>> What are you using for zkSpoutId?  Is it the same value every time you
>> submit the topology?
>>
>>
>> On Fri, Nov 6, 2015 at 4:38 PM, Birendra Kumar Singh <[email protected]>
>> wrote:
>> > Hi
>> >
>> > I am using Kafka with Storm as described here -
>> > https://github.com/apache/storm/tree/master/external/storm-kafka
>> > However, If I kill my topology and start it again, it restarts the
>> > KafkaSpout from offset 0. Ideally, it should restart from where it left
>> > off.
>> > My nimbus and supervisors are up and running all the time.
>> >
>> > I am using the default zookeeper provided by kafka.
>> >
>> > storm version - 0.95
>> > kafka - 2.9.1
>
>

Reply via email to